A Comprehensive Picture of Factors Affecting User Willingness to Use Mobile Health Applications

mHealth apps are being used in a variety of ways, such as disease prevention, health promotion, patient education, self-management, and remote monitoring [32]. Numerous studies have reported various benefits of mHealth apps, such as increased access to healthcare [55, 66], improved patient engagement [60], and more efficient healthcare delivery [17, 33]. Additionally, mHealth apps allow individuals to monitor their own health, track their progress, and receive personalized feedback and recommendations [6].

While mHealth apps offer many potential benefits, they also face several challenges. Studies by Baig et al. [10] and Newaz et al. [48] have identified security and privacy issues as major challenges facing mHealth apps. Jaime and colleagues [12] conducted a privacy assessment of mHealth apps based on 24 selected articles published from 2014 onwards. They found that despite great progress made in raising awareness of privacy and security in mHealth apps, there is still much to be done. Two major barriers identified were the diversity of mHealth apps and the lack of standard evaluation criteria. In addition, Tangari and colleagues [63] analyzed 20,991 mHealth apps on the Google Play Store and found serious problems with privacy and inconsistent privacy practices in mHealth apps. They urged clinicians to be aware of these issues and to articulate them to patients when determining the benefits and risks of using these apps.

While mHealth apps have great potential, some people may be hesitant to use them due to various reasons. These include concerns about the privacy and security of personal information [4, 78], a lack of trust in the accuracy of the information provided by the app, and concerns about the cost of the app [74].

Previous studies have indicated that the willingness to use mHealth apps can be affected by various factors, such as age, gender, health status, and technological ability [50, 70]. However, the results from these studies are not entirely conclusive. One study by Andreia et al. [50] found that older users exhibited a higher level of conscientiousness and behavioral intention to use mHealth apps. On the other hand, Virella et al. [70] reported that younger individuals were more likely to use mHealth apps than older adults. Gender also plays a role in the relationship between personality traits and the behavioral intention to use mHealth apps, according to Nunes et al. [50]. While Zhang et al. [82] found that males had a higher level of intention to adopt mHealth compared to females, a later study by Bol et al. [13] reported no gender effect on aggregated mHealth app use. Furthermore, Ernsting et al. [18] found that individuals with chronic conditions were more likely to use mHealth apps. However, the study by Robbins et al. [56] suggested that individuals with mHealth apps were not more likely to have chronic health conditions compared to those without.

The previous studies have generated inconsistent results. In our research, we expand on the previous research by examining a more extensive range of factors that affect the acceptance of mHealth apps, including those that have not been well researched, such as users’ cultural and educational backgrounds, as well as the data processing and storage of mHealth apps. Furthermore, we examine how users’ prior experience with Covid-19 affects their willingness to use mHealth apps.

In this subsection, we describe our research questions, and propose hypotheses under each research question.

RQ1: What are the factors that influence users’ willingness to use mHealth apps?

In this study, our goal is to present a comprehensive and unified perspective on the potential factors that impact user willingness to use mHealth apps. We propose that four broad categories of variables will influence this willingness. The first category is demographic background, including age, gender, and educational background. Second, user intention is influenced by perceived benefits and interest in the mHealth app, while perceived costs decrease willingness. Additionally, since mHealth apps handle users’ health data, privacy concerns and the risk of disclosure become more prominent. Therefore, trust in the data collection and sharing process and privacy concerns will be of significant importance to users. Finally, the willingness to use mHealth apps will be impacted by factors related to users’ technological background. These factors include user experience and comfort with mobile technology and feelings of control over the data management. We describe four hypotheses based on these factors below.

Previous studies have shown inconsistent results regarding the influence of age, gender, and health status on user willingness to use mHealth apps. While some studies have reported significant effects of these factors [18, 50, 70, 82], a few others have indicated limited effects [13, 56] (refer to Section 2.2 for a more detailed discussion). Recently, Utz and colleagues [65] reported that countries of residence also impact user preferences for the collection of personalized data or anonymity, with Chinese participants preferring the former and German and US participants favoring the latter. The variation in the impact of demographic background across studies may be partly due to different experimental settings and contexts. In our study, we hypothesize that if a demographic factor influences user views and preferences for mHealth apps, it is likely to impact user willingness to use mHealth apps. Based on this, we propose our first hypothesis:

H1: The willingness of users to use mHealth apps is influenced by their demographic background, including factors such as age, gender, educational background, and health status.

Previous research has shown the importance of perceived benefits and interest in user willingness of using mHealth apps [14, 75]. Therefore, our second hypothesis is:

H2: Users are more willing to use an mHealth app if they find it beneficial to their health, or if they can get financial rewards for using the app.

Privacy and security are paramount for any mobile app, particularly mHealth apps that deal with sensitive and personal information such as users’ health history, medication usage, and location data [71]. Prioritizing privacy and security in an mHealth app can increase user willingness to use it. A study by Zhou and colleagues [83] revealed that most study participants had concerns about their privacy when using mHealth apps and expressed their preferences for security features, such as regular password updates, remote wipe, user consent, and access control. Based on these findings, we propose the third hypothesis in this study:

H3: Users’ willingness to use an mHealth app is positively correlated with their trust in the app’s privacy and security measures.

The technological background of a user can refer to their experience with technology and their level of comfort in using it. Previous studies have suggested a positive correlation between users’ technological background and their willingness to use mHealth apps [41]. For instance, a study by Jaana and colleagues [29] revealed that older adults were less likely to use mobile health apps due to a lack of familiarity with technology. Thus, we propose the fourth hypothesis:

H4: Greater familiarity with technology is positively associated with user willingness to use mHealth apps.

RQ2: How do various factors jointly contribute to users’ intention to use mHealth apps?

Previous research suggests that multiple factors interact with each other in influencing users’ intention to use mHealth apps. For instance, gender has been found to moderate the effects of personality and mobile technology preferences, which, in turn, affect users’ willingness to use mHealth apps [50]. In our study, we will utilize structural equation modeling to quantitatively measure how various factors collectively contribute to users’ willingness to use mHealth apps, and draw a comprehensive picture of the underlying structure.

The experiment was approved by the Institutional Review Board (IRB) of National University of Singapore. A participation information sheet was shown at the first page of the experiment. Participants provided consent before starting the experiment.

We recruited a total of 1,669 participants (1,021 male, mean age 39.64 \(\pm\) 7.83) from eight countries (US, UK, Germany, India, China, Singapore, New Zealand, and Australia) across four continents. Among them, 436 participants were from the online crowd-sourcing platform Amazon Mechanical Turk (MTurk) [42], and 1,233 participants were from Toluna Global Panel, another crowd-sourcing platform managed by the Internet survey company Toluna (www.toluna-group.com). We used two platforms for two reasons: first, MTurk has few participants from Oceania and Asia (except India), so we used Toluna to obtain more participants from eight different countries; second, MTurk has few participants over the age of 60, but we aimed to recruit a significant number of seniors as they are a specific target audience of mHealth apps. Toluna enabled us to recruit seniors as well as participants in other age groups. The chosen countries span multiple continents, providing a diverse representation of global populations, cultures, and healthcare systems. For example, the US, UK, Germany, Australia, and New Zealand are known for their advanced healthcare infrastructure, making them valuable for studying app adoption and usage patterns. Among which, both UK and Germany are known for having strong privacy protection regulations and laws that are designed to safeguard individuals’ personal data and privacy rights. India and China represent rapidly growing mobile markets with large populations, allowing us to assess app utilization in regions experiencing technological and healthcare advancements. China also represents a substantial market for mobile health apps designed for Chinese speakers, considering that Chinese is one of the most widely spoken languages globally. Singapore is recognized for its strong technological infrastructure and government support for digital initiatives, offering insights into highly connected societies. Every country possesses a distinct healthcare system, spanning from public to private and market-driven models. This diversity provides an opportunity to examine how app usage corresponds with varying healthcare industry contexts. For each country, we attempted to achieve a uniform distribution of participants’ ages. Overall, our participants had diverse demographic backgrounds in terms of their residential countries, gender, and age (see Table 1 for details).

Our human experiment is based on the vignette design, which combines the advantages of survey research and multidimensional experimental design. In vignette experiments, short, systematically varied descriptions of situations or subjects (called vignettes) are used to elicit the beliefs, attitudes, or behaviors of respondents regarding presented scenarios. Participants evaluate hypothetical situations or subjects described in vignettes that vary in the level of characteristics (dimensions) of the described situations or subjects [34, 62]. The advantage of vignette experiments is that they allow researchers to study complex social phenomena in a controlled and systematic way. By using carefully constructed scenarios, researchers can manipulate different variables to examine their effects on people’s attitudes and behaviors. This can help isolate and identify specific factors that influence how people think and act in different social situations [25].

Our human study was performed on two crowd-sourcing platforms—Amazon Mechanical Turk (MTurk) and Toluna. We recruited participants from the US and India through MTurk and participants from other countries through Toluna. The vignettes and surveys were translated to Chinese for participants from China, while participants from other countries responded in English. Each participant was presented with eight vignettes, which described a hypothetical mHealth app. Participants rated the likelihood of using the app on a 7-point scale, where 1 indicated “Very unlikely” and 7 indicated “Very likely”. They were also asked to provide a reason for their response, rate the perceived usefulness of the app on a scale of 1 to 7, and indicate their major goal if they decided to use the app. Additionally, participants were asked about what they liked most and least about the app and their demographic information, such as gender, age, ethnicity, educational background, and country of residence. Questions were also included about participant information, such as digital literacy and online behavior. All measures used in the study are available in Appendix B.

We conducted statistical analyses using several methods to examine the impact of each independent variable on the dependent variables. First, we conducted ANOVA (analysis of variance) to check for any overall effects. If ANOVA results were significant, we performed post-hoc Tukey tests to identify specific manipulation effects. For analyses involving multiple factors, we used both logistic regression and linear regression. To control for multiple comparisons, we applied the Bonferroni correction. For more information about these inferential statistics, please refer to Bailey’s guide [11].

We used structural equation modeling to gain a comprehensive understanding of how multiple factors collectively influence users’ behavioral intention. In particular, we first conducted an exploratory factor analysis (EFA) to identify the number of latent variables (factors) present in the human data and extract a concise set of attributes underlying these variables. Next, we conducted a confirmatory factor analysis (CFA) to assess the reliability and validity of the measures and test the relations between latent and observed variables. Factor analysis (FA) and principal component analysis (PCA) are variable reduction techniques that extract a reduced set of variables from a larger set of variables, but in PCA, the components are orthogonal linear combinations that maximize the total variance, while in FA, the factors are linear combinations that maximize the shared portion of the variance underlying “latent constructs.” Attributes with poor loadings or fits were removed. We developed the final model through path analysis, which calculated standardized regression weights ( \(\gamma\) ) and correlations ( \(\phi\) ) among latent variables. We conducted these analyses using statistical toolboxes in Matlab R2016b, IBM Amos 21, and IBM SPSS 20.

Readers can view the vignettes and try our human study online at https://ncript.comp.nus.edu.sg/site/experiment2/start?taskid=4391.

We conducted ANOVAs for the seven factors listed in Table 2 to examine whether each factor influences participants’ willingness and perceived app usefulness. Results indicated that three factors (rewards for using the app, the types of data collected by the app, and how data is shared by the app) significantly impacted participants’ willingness, with Fs \(\ge 5.13\) and ps \(\le .006.\) ¹ Financial rewards (e.g., discount/coupons for health service and products) had greater effect on promoting user willingness than knowing and interacting with people with similar health conditions, with \(F(2,14518) = 5.13\) and \(p = .006\) . The factor of how data is stored had a marginal impact on participants’ willingness, with \(F(2,14518) = 5.01\) and \(p = .007\) . App functionality, how the app data is protected, and users’ level of control did not have significant impact. Only one factor, the types of data collected by the app, had a significant impact on perceived app usefulness, with \(F(2,14518) = 26.13\) and \(p \le .0001\) . Detailed post-hoc Tukey test results for the impact of separate factor levels on participants’ willingness are provided in Table 3.

The ANOVA revealed interaction effects of country of residence with app rewards and data collection on users’ willingness, with Fs \(\ge 2.16\) and ps \(\le .007\) . Follow-up analyses showed that the effect of app rewards was primarily driven by German participants ( \(F(2,1004) = 5.23\) , \(p = .006\) ) and New Zealand participants ( \(F(2,1467) = 5.17\) , \(p = .006\) ), while the effect of data collection was significant among most countries except India, China, and the United Kingdom, with Fs \(\ge 6.78\) and ps \(\le .001\) . No interaction effect of data collection and country of residence was found for perceived app usefulness.

In this section, we will explore the impact of participants’ demographic backgrounds on their willingness to use the app and their perceived usefulness of it. We will be focusing on five key demographic factors across all app scenarios: country of residence, gender, age, ethnicity, and educational background.

The ANOVA indicated a significant effect of country of residence on both user willingness to use mHealth apps ( \(F(7,14513) = 116.08\) , \(p \le .0001\) ) and perceived app usefulness ( \(F(7,14513) = 147.29\) , \(p \le .0001\) ). Indian participants had the highest willingness to use the apps (ps \(\le .001\) ), followed by Chinese and American participants. Participants from Germany and New Zealand showed the lowest intention to use mHealth apps (ps \(\le .05\) ). Participants’ perceived usefulness of the app followed a similar pattern (see Figure 2).

The age of the participants had a significant impact on both their willingness to use the mHealth apps ( \(F(7,14515) = 387.64\) , p \(\le .0001\) ) and their perceived usefulness of the app ( \(F(7,14515) = 132.63\) , p \(\le .0001\) ), as illustrated in Figure 3(a). The distribution of age over user willingness formed a bell curve, with participants in the age group of 25–34 exhibiting the highest willingness to use the app and perceiving it to be the most useful (ps \(\le .001\) ). The second-highest willingness and perceived usefulness were demonstrated by the age group of 35–44 (ps \(\le .001\) ). Conversely, participants aged below 19 and above 60 showed the lowest willingness to use the app and perceived its usefulness to be the least (ps \(\le .05\) ).

Notably, participants in the age group of 45–59 reported a similar perceived usefulness of the app as those in the age group of 19–24, but the former group demonstrated a lower willingness to use the app compared to the latter. Additionally, participants aged 35–44 rated the app’s usefulness higher than those in the age group of 19–24, yet both groups exhibited statistically similar intention to use the mHealth app. These findings suggest that other factors, such as technical skills and related experience, may influence users’ behavioral intention. Therefore, a more comprehensive analysis is necessary, as discussed later in Section 4.4.

Additional influential factors include participants’ educational background and ethnicity. Participants’ level of education profoundly influenced their willingness of using mHealth apps ( \(F(4,14516) = 241.20\) , p \(\le .0001\) ) as well as their perceived app usefulness ( \(F(4,14516) = 182.49\) , p \(\le .0001\) ). Notably, a near-linear correlation emerges between increasing education and heightened user willingness, as depicted in Figure 3(b). This observation was reinforced by post-hoc Tukey tests (ps \(\le .05\) ). Participants’ ethnicity also impacted users’ willingness ( \(F(5,14515) = 94.27\) , p \(\le .0001\) ) and perceived app usefulness ( \(F(5,14515) = 87.49\) , p \(\le .0001\) ). Asians (including Asian Indians) and American Indians showed greater willingness to use the app than Black people (ps \(\le .001\) ), while White and Hispanic participants were the least willing (ps \(\le .05\) ). Gender only had a marginal influence on users’ willingness to use the app ( \(F(1,14519) = 3.25\) , p \(= .071\) , male: .68, female: .67), but profoundly impacted users’ perceived usefulness of the app ( \(F(1,14519) = 23.99\) , p \(\le .0001\) ). Male participants perceived the app as more useful than female participants (male: .65, female: .63, \(t(14519) = 4.90\) , p \(\lt .0001\) ).

We conducted exploratory analyses on individual difference measures to determine which factors predicted users’ willingness to use the mHealth apps. To do this, we calculated the average willingness of each participant to use the eight different versions of the mHealth apps described in the vignettes. We then performed a multiple linear regression analysis using the participants’ willingness as the dependent variable and individual difference measures as explanatory variables.

Table 4 shows that variables related to privacy and security concerns, online behaviors, prior experience of related apps and the Covid-19 pandemic, and technology skills robustly impacted users’ willingness to use mHealth apps. The strongest predictor was users’ smartphone skills, followed by users’ positive attitude towards wearable devices, and users’ frequency of sharing health information online. Surprisingly, whether a user had a chronic disease or whether they had used mHealth apps before did not have a significant impact.

In this section, we utilized structural equation modeling (SEM) to gain a comprehensive understanding of how users’ behavioral intention towards our mHealth apps is influenced by various factors. While our previous analysis using multiple linear regression provided insight into the individual impact of variables, SEM allows us to explore the intercorrelations among these variables and how they jointly shape user willingness.

To begin, we conducted an EFA followed by a CFA [36] to measure the relationships between observed variables and latent factors (higher-level perceptions and reactions). We hypothesized that the reasons for user willingness are multidimensional and inter-correlated, so we used maximum likelihood with oblique transformation in our EFA [35]. Finally, we employed path analysis to inform the model structure.

The resulting model, shown in Figure 4, is divided into three layers. The individual difference measures obtained from our questionnaire are at the bottom layer (in grey rectangles) and load onto different latent variables represented by blue eclipses in the mid-layer. The top layer contains the latent variable “behavioral intention to use mHealth apps”, represented by an orange eclipse. We created the final model through path analysis, predicting the top layer latent construct from the lower-level perception latent constructs.

We evaluated the model’s fitness using two metrics: the Comparative Fit Index (CFI) and Root Mean Square Error of Approximation (RMSEA). Our model had acceptable fit, with \(CFI = .983\) and \(RMSEA = . 045\) . CFI compares the target model’s chi-square to an independent model, while RMSEA estimates the error of approximation per model degree of freedom and considers the sample size. Higher CFI and lower RMSEA values indicate better model fit.

In our SEM analysis, as depicted in Figure 4, five latent factors (represented by blue eclipses) were found to jointly contribute to users’ behavioral intention to use mHealth apps. Among these, “digital literacy” had the strongest weight on users’ behavioral intention ( \(\gamma = .44\) ), followed by “Online behavior of sharing personal information” ( \(\gamma = .25\) ), “Indifference to personal privacy” ( \(\gamma = .18\) ), and “Concern about personal privacy” ( \(\gamma = -.14\) ). The factor “Covid-19 related experience” had the lowest weight on users’ behavioral intention ( \(\gamma = .07\) ).

The individual difference measures in our study were moderated by users’ demographic background, including age, gender, ethnicity, education, and country of residence. As discussed in Section 4.1, we found significant interaction effects of country of residence with app rewards and data collection. Gender also interacted with several measures on users’ willingness, such as “having good cell phone skills” (gender: \(F(6,14507) = 18.61\) , \(p \le .0001\) ) and “concern about personal info leakage via mHealth apps” (gender: \(F(4,14511) = 7.28\) , \(p \le .0001\) ). Specifically, the impact of “having good cell phone skills” on users’ behavioral intention had a stronger effect on female participants ( \(\eta _{p} = .21\) ) than male participants ( \(\eta _{p} = .11\) ). In contrast, the influence of “concern about personal info leakage via mHealth apps” was carried by male participants ( \(F(4,8786) = 18.61\) , \(p \le .0001\) ), with no significance found among female participants ( \(F(4,5725) = 1.23\) , \(p = .296\) ).

Furthermore, we observed a significant interaction effect between education and “being tech-savvy” on users’ willingness ( \(F(16,14496) = 13.79\) , \(p \le .0001\) ). For example, the impact of “being tech-savvy” had a stronger effect size for participants with education below high school ( \(\eta _{p} = .64\) ) than for those with a graduate degree ( \(\eta _{p} = .21\) ).

In this section, we aim to investigate participants’ usage purposes and preferences for mHealth apps to gain a better understanding of users’ behavioral intention. The majority of participants selected “personal health monitoring” as their primary goal for using the proposed mHealth apps, followed by “getting personalized feedback from the app” and “receiving advice from health professionals”. Conversely, the least chosen option was “getting user rewards of using the app” (see Figure 5). This trend was consistent across all eight countries.

To gain further insight into participants’ opinions, we asked them to identify what they liked most and least about the proposed app separately. As depicted in Figure 6, participants appreciated the function of “personal health monitoring & disease prevention” the most and expressed their dislike for data sharing with other parties. Participants showed relative indifference to “users’ level of control over the app” and “user rewards for using the app” in both questions. This trend was consistent across all countries, with the exception of Chinese participants, who indicated that their least favorite feature was privacy protection instead of data sharing with other parties. Participants also provided reasons for their motivation to use or not to use the app, which were consistent with their preferred and least preferred features of the proposed app. For further details, readers can refer to Appendix A.

Our findings suggest that the development of users’ behavioral intention to use mHealth apps has a complex and hierarchical underlying structure. Multiple factors, such as users’ demographic background and technology skills, and the design of the mHealth app, jointly influence users’ willingness.

Our analyses showed significant impact for three app related factors: the rewards of using the app, types of data collected by the app, and with whom the data is shared. Our participants cared more about financial rewards (e.g., discount/coupons for health service and products) than knowing and interacting with people with similar health conditions. The effect of financial incentives is consistent with previous studies in [64]. However, our research only analyzed short-term acceptance. In terms of continued usage of mHealth apps, other factors may came into play, such as performance expectancy and price value [75].

Our study found that participants were more willing to allow mHealth apps to collect their health records, such as medication and clinic consultation records, than their daily usage data on mobile phones, such as the number of messages and phone calls, browser histories, and app usages. This preference may be due to the clear linkage between health records and mHealth apps, while the connection between the app and daily usage on mobile phones is unclear. However, previous studies have shown that daily phone usage can be informative of a person’s mental health [30, 31, 45]. We anticipate challenges in raising user awareness and acceptance of mHealth apps that are related to mental health.

Furthermore, our study found that participants’ intention to use mHealth apps was negatively influenced by sharing app data with third parties such as app developers and hospitals. Previous research has identified multiple factors that influence users’ willingness to share self-collected health data, including the source and type of data, user benefits of data sharing (such as personalized feedback), and privacy and security concerns [23, 39, 73]. To encourage the sharing of mobile health data for the benefit of individuals and the wider community, greater efforts are needed to address these concerns and promote the benefits of data sharing.

Our analysis did not find a significant influence of app functionality, but this does not necessarily imply that functionalities lack significance. It is possible that the four levels on our functionality dimension are not sufficient to represent the range of functionality offered by mHealth apps in general. Further research should consider more detailed and specialized functionality features that cater to specific groups, such as pregnancy apps designed for women. Such specialized functionality may have a greater impact on users’ behavioral intentions to use mHealth apps than general functionality measures. In addition, past studies have indicated that there is a significant correlation between higher levels of perceived ease of use and the intention to use a product or service [69]. Our forthcoming research will also explore the significance of user-friendly attributes as a crucial factor in this regard.

Our analysis revealed that users’ willingness to use mHealth apps was significantly influenced by their demographic features, such as country of residence, age, and education.

Participants from India, China, and the United States exhibited the highest intention to use the apps, with Singapore following closely behind. In contrast, participants from Europe (United Kingdom and Germany) and Oceania (Australia and New Zealand) showed relatively lower intention compared to those from Asia and the United States. Moreover, the intention of participants from Germany and New Zealand was influenced by app rewards, while those from other countries were not. The type of data collected by the app also significantly influenced the intention of participants from most countries, except India, China, and the United Kingdom.

Several factors could account for these differences. Cultural attitudes towards healthcare and data sharing, for instance, may vary across regions, with some cultures emphasizing self-care and being more open to data collection and sharing. For instance, research by Utz and colleagues [65] found that Chinese participants preferred the collection of personalized data, while German and US participants favored anonymity. In some regions, such as India and China, limited access to healthcare services in certain regions could make mHealth apps a more attractive option for individuals seeking to manage their health [53, 77]. Differences in technology adoption and digital literacy could also play a role, with some populations being more comfortable with using mobile devices and apps in their daily lives [46]. Finally, varying regulatory frameworks and policies related to mobile health apps could also affect acceptance levels [8, 81].

Different from previous research [50, 70], our study identified a bell curve pattern that linked user age and their willingness to use mHealth apps. We found that participants aged between 25 and 34 expressed the strongest intention to use these apps, followed by the age group of 35–44. Participants over the age of 45 or under the age of 25 demonstrated a lower behavioral intention. In particular, participants above the age of 60 reported the lowest willingness to use the apps. Education enhances users’ willingness to use mHealth apps. Our SEM analyses suggest that one possible explanation for this trend is that individuals with higher levels of education may possess greater digital literacy.

Our study provides strong evidence that several demographic factors, such as country of residence, age, gender, and education, play a significant role in moderating mHealth adoption. While previous literature has highlighted age and gender as commonly recognized moderators [27, 49], our research offers more nuanced insights by revealing additional moderating factors, including education and country of residence.

Our research may have relevance across various disciplines. First of all, our findings can guide mHealth developers in crafting apps that resonate with users. We suggest designers take into account influential factors on users’ willingness to adopt the app, such as the type of data collected by the app and its sharing recipients. Offering users options for diverse data collection, along with clarifications of each data’s benefits, is recommended. Providing transparent control over data sharing and ensuring flexibility is crucial. Additionally, users are interested in app rewards and prioritize them over data protection and storage concerns. mHealth service providers can emphasize these rewards to boost service utilization.

The most significant finding of our SEM analyses is the critical role of users’ digital literacy. As the reliance on digital tools increases, it has the potential to exacerbate existing health disparities by widening the gap between those who possess digital skills and access to digital tools and those who do not. Considering the pivotal role of users’ digital literacy in ensuring equitable access to mHealth apps, governments should implement comprehensive digital literacy programs that empower citizens with the skills to effectively use and navigate digital tools, fostering inclusivity and bridging the digital divide. For example, governments can forge partnerships with health organizations and app developers to devise specialized educational initiatives for mHealth technology. Additionally, they can champion the idea of lifelong learning by stressing that digital literacy is a skill attainable at any age. In regions characterized by low digital literacy rates, such as developing countries, governments can collaborate with non-governmental organizations (NGOs), international development agencies, educational institutions, and private sector entities to pool resources and expertise in promoting digital literacy. Given the importance of data collection and sharing highlighted in the previous paragraph, our research also suggests that governments should establish clear regulations and guidelines for data management. Governments can advocate for standardized privacy settings and data sharing consent mechanisms to ensure users have meaningful choices over their personal information.

Our study also offers valuable insights for healthcare professionals and marketers aiming to advance and integrate mHealth interventions. Through showcasing the notable moderating impacts of demographics, our research underscores the necessity of customizing digital health solutions for diverse populations, encompassing an array of demographic factors. For instance, the demand for specialized apps tailored to seniors with limited digital literacy is particularly evident.

Finally, our study also presents an invitation for researchers to delve deeper into the underlying structure that shapes users’ behavioral intention to use mHealth apps. Through this exploration, we hope to gain a better understanding of how different factors, such as users’ demographics and online behavior, jointly impact user willingness to utilize mHealth apps. This deeper understanding will allow for more comprehensive explanations of user behavior and pave the way for improved mHealth app development and adoption.